Background
We aimed to evaluate the feasibility of conducting a randomized trial of a task-oriented, community-based exercise program incorporating a healthcare-recreation partnership called Together in Movement and Exercise (TIMETM) to improve function post-stroke. 

Methods
In an ongoing pilot trial, 33 community-dwelling individuals able to walk 10 metres independently post-stroke were stratified by gait speed and randomized to the TIMETM program (n=16) or a wait-list control group (n=17). TIMETM involves repetitive, progressive practice of functional balance and mobility tasks (2 classes/week, 12 weeks). Of blinded evaluations scheduled at 0, 3, 6, and 12 months, 0- and 3-month evaluations have been completed. 

Results
At 2 hospitals, a stroke rehabilitation professional recruited 17/46 (37%) and 9/50 (18%) individuals in-person and/or by phone. At 1 hospital, a research assistant recruited 7/358 (2%) through mailings/calls. Six participants (38%) assigned the TIMETM program attended <50% of classes citing new (melanoma, hip fracture, suspected stroke) or exacerbated existing health conditions (foot edema/blistering, back pain, fatigue). Two (13%) people assigned the TIMETM program withdrew or were lost to follow-up due to back pain (n=1) and inability to contact (n=1). Three people assigned usual care (18%) did not want to wait 12 months for the TIMETM program; 2 (12%) withdrew. Four of the 8 fitness instructors (50%) that physiotherapists trained and supervised to deliver TIMETM in recreation centres became unavailable necessitating additional recruitment and training. 

Conclusions
Protocol recommendations include: engage stroke team members as recruiters; screen extensively for health conditions impacting exercise participation; reduce control-group wait-time; and train 4 instructors/centre to optimize availability.

FEASIBILITY OF A RANDOMIZED TRIAL EVALUATING A GROUP, TASK-ORIENTED COMMUNITY EXERCISE PROGRAM INCORPORATING A HEALTHCARE-RECREATION PARTNERSHIP FOR PEOPLE WITH MOBILITY LIMITATIONS POST-STROKE